[Qemu-devel] QEMU CVS, X-Code 2.0 and Tiger

[Qemu-devel] QEMU CVS, X-Code 2.0 and Tiger

[Natalia Portillo]

Hi!

I'm getting (along with a LOT of warnings) the following error when  
trying to compiler under Tiger/X-Code2 (GCC 4.0.0-apple):

TiBook:~/qemu claunia$ make
for d in i386-softmmu ppc-softmmu sparc-softmmu x86_64-softmmu; do \
make -C $d all || exit 1 ; \
         done
../dyngen -c -o opc.h op.o
dyngen: blr expected at the end of op_bsfl_T0_cc
make[1]: *** [opc.h] Error 1
make: *** [all] Error 1

Any idea?

Regards

Re: [Qemu-devel] QEMU CVS, X-Code 2.0 and Tiger

[Paul Brook]

On Wednesday 04 May 2005 07:48, Natalia Portillo wrote:
> Hi!
>
> I'm getting (along with a LOT of warnings) the following error when
> trying to compiler under Tiger/X-Code2 (GCC 4.0.0-apple):
>
> TiBook:~/qemu claunia$ make
> for d in i386-softmmu ppc-softmmu sparc-softmmu x86_64-softmmu; do \
> make -C $d all || exit 1 ; \
>          done
> ../dyngen -c -o opc.h op.o
> dyngen: blr expected at the end of op_bsfl_T0_cc
> make[1]: *** [opc.h] Error 1
> make: *** [all] Error 1
>
> Any idea?

Qemu doesn't work with gcc4.

Qemu makes assumptions about the structure of the code that gcc generates. 
These assumptions are no longer valid when compiled with gcc4.

Paul

Re: [Qemu-devel] QEMU CVS, X-Code 2.0 and Tiger

[Pierre d'Herbemont]

On 4 mai 05, at 08:48, Natalia Portillo wrote:

> Hi!
>
> I'm getting (along with a LOT of warnings) the following error when  
> trying to compiler under Tiger/X-Code2 (GCC 4.0.0-apple):
>
> TiBook:~/qemu claunia$ make
> for d in i386-softmmu ppc-softmmu sparc-softmmu x86_64-softmmu; do \
> make -C $d all || exit 1 ; \
>         done
> ../dyngen -c -o opc.h op.o
> dyngen: blr expected at the end of op_bsfl_T0_cc
> make[1]: *** [opc.h] Error 1
> make: *** [all] Error 1
>
> Any idea?

switch to gcc-3.3... It is still included with in the package. Or use  
Daniel Egger solution (seems to work only for i386-softmmu):

On Thu, 14 Oct 2004 12:09:45 +0200, Daniel Egger <degger@fhm.edu> wrote:

> I just wanted to mention to those of you trying to compile Qemu
> on Darwin with the latest gcc snapshots that you need to supply
> the -fno-tree-ch flag for the compilation of op.c at
> certain optimization levels for dyngen to work[1]. Performance
> of i386-softemu qemu is quite descent, and the resulting code
> is smaller than with the native gcc 3.3 but the ppc emu doesn't
> compile.
>
> [1] Otherwise there'll be no branch to link register (blr)
>      instruction at the end of some of the virtualized
>      instructions which dyngen relies on to find the borders.
>


Pierre.

Re: [Qemu-devel] QEMU CVS, X-Code 2.0 and Tiger

[Daniel Egger]

[-- Attachment #1: Type: text/plain, Size: 837 bytes --]

On 05.05.2005, at 10:09, Pierre d'Herbemont wrote:

> switch to gcc-3.3... It is still included with in the package. Or  
> use Daniel Egger solution (seems to work only for i386-softmmu):

>> I just wanted to mention to those of you trying to compile Qemu
>> on Darwin with the latest gcc snapshots that you need to supply
>> the -fno-tree-ch flag for the compilation of op.c at
>> certain optimization levels for dyngen to work[1]. Performance
>> of i386-softemu qemu is quite descent, and the resulting code
>> is smaller than with the native gcc 3.3 but the ppc emu doesn't
>> compile.

In fact I've just verified that -fno-tree-ch still does the trick
with the Tiger gcc 4.0.

I still don't care about the other emus (and also have no chance
to test them) so it might work but no guarantees from my side. ;)

Servus,
       Daniel


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 186 bytes --]